Geomorphological understanding of floods

Abstract

The summer 1993 flooding of the upper Mississippi River valley reminds us that floods are the most globally pervasive, environmentally diverse and continually destructive of all natural hazards. The fact that flood damages continue to rise raises commonsense questions about conventional flood science. Like much modern environmental science, conventional flood science has followed the model of theoretical physics. It advanced from early emphasis on streamflow measurement to the use of simple formulae, and finally to the abstract theoretical sophistication of modern modeling studies. Two approaches are now used to “predict” flood phenomena: (1) beginning with the conventional database of measured properties of small common floods, a conceptual generalization is made to the idealized properties of the large, rare floods from which society is assumed to be at risk, and (2) explanation of detailed, specific flood phenomena is achieved through theoretical generalization (models) based on “first principles”, which are assumed to apply to the entire class of phenomena. Unfortunately, both approaches devote almost all their attention to methodology, increasingly mathematical, without questioning basic underlying assumptions. Increasingly it is the assumptions, often unstated, that serve to embody the understanding of floods as real-world particular phenomena, rather than as conceptual generalities. Such trends lead to an unease that it is not floods that are being researched by much of conventional flood science. Rather, such flood “science” is increasingly becoming the mathematical manipulation of idealized parameters that are assumed to have flood-like properties. These idealizations of flood attributes are generalized, and the resulting predicted consequences are imposed upon society through engineering designs, flood-hazard zonations, and the like.